It is 25$ + server cost for 500 CCU. If you got around 500 CCU constantly you have a player-base around 200K active monthly users. If you got that the price is no problem at all. And until you get more than 100 CCU active users its actually free!
It does got a steep learning curve if your never worked with something like it before or you are a hobbyist programmer. But when you get into it it is pretty simple. Of cause not as simple as Photon Cloud, but this is a tool for people who are serious with their games. It is SO powerful and can be used for all types of games. It’s been used for fast paced FPS to MMO’s, its proven, stable and got features that lets you easily expand once you got success.
But yes it is more work as you need do your own server side collision and more (there is help out there for that), but then you don’t have to run each game in a Unity Instance, so with a lot of users it is more cost effective.
Enough rambling about photon! People might start to believe I work for them haha
Actually, Unity’s physics is highly optimized (probably more than anything an independent developer can create) and you can easily start Unity instances headlessly (accepting no user input and presenting no graphics) which is just as efficient as running Photon’s load balancing.
I’ve actually started using Forge for my project. Upon further exploration, I found that it has just as much flexibility as DarkRift as it’s event-driven. Performance is of no concern; simple server logic results in no more than .6% of my CPU usage.
Well the good news is that we are going to be offering “Forge Bare Metal” in the beta as well. This is a server that runs without a Unity instance so that if you are doing a pure messaging based game/application and do not need anything from Unity, you can use this no-overhead solution
Clearly you have not worked with Photon Server. A game(room) is just a class instance that run all logic. You can run 100’s of games on one server. Load-balancing is not important before you have more than 100’s of CCU and need to have multiple servers. Loadbalancing is for sending people to new servers when one of your servers is full, so you can just open new servers as you grow.
And i know for a fact that a class instance is not as heavy as a headless instance for each game. And i played with Forge too, and already started creating my own load balancing. But thats not something it got out of the box.
Unity’s physics is PhysX and is not even at the newest version. Its getting updated in 5.0 though.
I would like to note the scaffolding for this is in place and will find it’s way into the system very soon
We needed it for something we were working on ourselves and we saw that others were wanting this flexibility too. So, as we have always done, we work to make the system as powerful, flexible and easy to use as possible from as many perspectives as possible
I’m not sure why you assume I haven’t worked with Photon because I have - I had my entire lockstep framework set up on Photon at one point. There’s little benefit to using Photon over other, more efficient networking solutions. When you can run 100+ instances on a single machine, the differences become negligible (I don’t think I have even enough bandwidth to support that many instances).
Even so, I’m very glad there will be a bare bones version. This is more proof of the capability of Forge’s developers as well as matching (if not exceeding) other networking solutions.
My main point is that it would be more efficient to do simulations in a headless Unity instance than with a custom physics library (painfully) integrated into the server SDK. Besides, authoritative server attempts with Photon by both developers and the authors have failed so there’s you don’t even know if it can work. Photon definitely has its uses, but it can’t be something it’s not.
As for the 100CCU, it is useless except for trial (thank goodness). To have over 500 CCU which many games need, the developer will has to cough up $3500. Their tiers of licenses seems like a ploy to take money from any game unfortunate enough to be successful.
i am not sure about that, maybe i did not read correctly, but once i read in a few articles of existing mmos and their experience with users and most of them mention that usually 10% of your userbase are connected at any gicen time,
so if this stas are correct then 10% of 200k users are around 20k ccu online.
not 500.
i may missread the mmo articles or got the numbers confused, but so far that is how i understood.
so if one game has around 500 ccu at any given time, then it may mean that the total userbase count might be up to 5k.
like i said i may be wrong, so dont belive what i thought i read about mmo stats articles.
At their record peak, and this is at a single occasion they at a bit below 15% online.
But lets say your game for some reason got a constant 20% CCU of the player-base on then you need to sell 2500 games. Now lets say you earn 10 dollars per game then thats 25000$. You can now buy a unlimited CCU license and still keep 21500$ for wages, server cost and so on.
Now if you used the free 100 CCU license then you need to sell at-least 500 copies to fill that. You would get enough money using the free license to buy the most expensive one they got.
But do you think that if you had 500 players then 24 hours a day at-least 100 would be on at all times? Thats highly unrealistic. So you could probably make a lot more money using the free license.
I think you are a bit off on your calculations,
none of us can not compete against bigdogs studios like arenanet, blizzard etc,…
we are being realistic, we are independent game developers ,must of us are not even a team of 2,
therefore fundings are coming out of our own pockets from our primary day job.
we all share a nice dream and we all use what is at our reach,
if one of us releases a playable multiplayer game, our only hope for succeed is to offer it free to play,
100 or 500 and even 10k can take place overnight if your game is a little bit decent.
tell me,
how you gonna afford a 4k license fee for photon with your paycheck,
you still have to feed your kids and wife.
so, what i am saying, maybe you do have the resources to buy a photon license.
but not most of us, or at least myself.
hence we all gother here and get exited on what we can afford and what others can offer that fits our pocket range and our planned project.
we know you like photon and i am happy for you,
but we want to move on forward to our needs and availability to reach them…
so thank you again for being an ambassador of photon, but we all used it and most of us would like to put that behind and find our one that best fits our needs and budget.
I created a game that will be on steam in a not to distance future. Its multiplayer but not free to play (very cheap though), so lets hope i can get some buys anyways
I would say free to play is a lot bigger gamble than paid niche games. If you create a free to play game that gains success overnight, but your in-game shop does not work (wrong design) then your can come in huge financial problems because you will suddenly get huge server (bandwidth) bills! With paid games at least u won’t get that traffic before people paid. Overall if you have success then networked licenses is not your problem, server costs are.
Actually worked a bit as a volunteer programmer at a company a few years back, on a free to play game to get some CV. It did not succeed but did in a short period have an okay amount of players online. Trust me, the server costs can be insane
That looks like a fun game! Great job! I guess I was being a bit harsh on Photon mainly because my Photon+ purchase isn’t worth anything to me. They do have their merits.
Ah, that’s very true. I can imagine how prototyping a game with Photon would be super easy. Just drag and drop that networkview and observe the transform.
exactly, #1 reason to avoid photon, very large BW overhead.
I sure would not like to pay 3,800 usd for 1 server license for photon up in front,
BW cost are my main concern so dont want extra 4k in top of that to worry about.
no need to trust you, been managing them for the past 14 years (web servers from webhosting company i work for)
even tho too much difference between web and game servers but prices for BW were excruciating back in the day.
hence i know how much it takes to run a game server , which has forced me to stop using photon and run as fast as i could and never look back.
Actually not weird claims, to me are valid and tangible claims that for my game project will suck me dry before i even start my first couple of close beta testing days.
unless you willing to pitch in with a good donation to my project, otherwise i need to get my feet as close to the ground as possible to even get a shot at the product launch.
you mention you been there so i assume you know what it takes, to get a shot at it and trying to not fail without even starting.
again, getting rid of photon was my first step to not fail, so now i got to think step to not fail before even start.
Note: I made the reviews (especially for Forge and DarkRift) more up to date with what I know about them and removed the PlayFab review because it’s not a networking solution.
But this BW claim is based on what? Yes their RPC’s in Photon Cloud has some overhead, but as far as a know thats not a problem with the Photon Server Solution. It is this claim that has annoyed me and why i defended them.
